Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Emerging Freight Truck Technologies: Effects on Relative Freight CostsFinal ReportMay 2018
Sponsored byMidwest Transportation CenterU.S. Department of Transportation Office of the Assistant Secretary for Research and Technology
About MTCThe Midwest Transportation Center (MTC) is a regional University Transportation Center (UTC) sponsored by the U.S. Department of Transportation Office of the Assistant Secretary for Research and Technology (USDOT/OST-R). The mission of the UTC program is to advance U.S. technology and expertise in the many disciplines comprising transportation through the mechanisms of education, research, and technology transfer at university-based centers of excellence. Iowa State University, through its Institute for Transportation (InTrans), is the MTC lead institution.
About InTransThe mission of the Institute for Transportation (InTrans) at Iowa State University is to develop and implement innovative methods, materials, and technologies for improving transportation efficiency, safety, reliability, and sustainability while improving the learning environment of students, faculty, and staff in transportation-related fields.
ISU Non-Discrimination Statement Iowa State University does not discriminate on the basis of race, color, age, ethnicity, religion, national origin, pregnancy, sexual orientation, gender identity, genetic information, sex, marital status, disability, or status as a U.S. veteran. Inquiries regarding non-discrimination policies may be directed to Office of Equal Opportunity, 3410 Beardshear Hall, 515 Morrill Road, Ames, Iowa 50011, Tel. 515-294-7612, Hotline: 515-294-1222, email [email protected].
NoticeThe contents of this report reflect the views of the authors, who are responsible for the facts and the accuracy of the information presented herein. The opinions, findings and conclusions expressed in this publication are those of the authors and not necessarily those of the sponsors.
This document is disseminated under the sponsorship of the U.S. DOT UTC program in the interest of information exchange. The U.S. Government assumes no liability for the use of the information contained in this document. This report does not constitute a standard, specification, or regulation.
The U.S. Government does not endorse products or manufacturers. If trademarks or manufacturers’ names appear in this report, it is only because they are considered essential to the objective of the document.
Quality Assurance StatementThe Federal Highway Administration (FHWA) provides high-quality information to serve Government, industry, and the public in a manner that promotes public understanding. Standards and policies are used to ensure and maximize the quality, objectivity, utility, and integrity of its information. The FHWA periodically reviews quality issues and adjusts its programs and processes to ensure continuous quality improvement.
Technical Report Documentation Page
1. Report No. 2. Government Accession No. 3. Recipient’s Catalog No.
4. Title and Subtitle 5. Report Date
Emerging Freight Truck Technologies: Effects on Relative Freight Costs May 2018
6. Performing Organization Code
7. Author(s) 8. Performing Organization Report No.
Ken Bao and Ray A. Mundy
9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)
Center for Transportation Studies
University of Missouri–St. Louis
240 JC Penny North, One University Boulevard
St. Louis, MO 63121-4400
11. Contract or Grant No.
DTRT13-G-UTC37
12. Sponsoring Organization Name and Address 13. Type of Report and Period Covered
Midwest Transportation Center
2711 S. Loop Drive, Suite 4700
Ames, IA 50010-8664
University of Missouri–St. Louis
One University Boulevard
St. Louis, MO 63121-4400
U.S. Department of Transportation
Office of the Assistant Secretary for
Research and Technology
1200 New Jersey Avenue, SE
Washington, DC 20590
Final Report
14. Sponsoring Agency Code
15. Supplementary Notes
Visit www.intrans.iastate.edu for color pdfs of this and other research reports.
16. Abstract
This research sought to evaluate the broad impacts that automated and connected vehicle technologies can have on both the motor
carrier and rail industries. Since the development and adoption of these technologies are likely to be gradual, three phases were
posited and analyzed. Depending on the degree of autonomy that is available, the motor carrier industry could achieve up to a
42.1% reduction in average cost per mile. And if fully autonomous technology was made available for use in the motor carrier industry, it is estimated that the American rail freight industry could see a 19% to 45% drop in demand.
17. Key Words 18. Distribution Statement
autonomous vehicle—connected vehicle—cost impacts—platoon—rail impact No restrictions.
19. Security Classification (of this
report)
20. Security Classification (of this
page)
21. No. of Pages 22. Price
Unclassified. Unclassified. 26 NA
Form DOT F 1700.7 (8-72) Reproduction of completed page authorized
EMERGING FREIGHT TRUCK TECHNOLOGIES:
EFFECTS ON RELATIVE FREIGHT COSTS
Final Report
May 2018
Principal Investigator
Ray A. Mundy, John Barriger III Professor for Transportation Studies and Director
Center for Transportation Studies, University of Missouri–St. Louis
Research Assistant
Ken Bao
Authors
Ken Bao and Ray A. Mundy
Sponsored by
University of Missouri–St. Louis,
Midwest Transportation Center, and
U.S. Department of Transportation
Office of the Assistant Secretary for Research and Technology
A report from
Institute for Transportation
Iowa State University
2711 South Loop Drive, Suite 4700
Ames, IA 50010-8664
Phone: 515-294-8103 / Fax: 515-294-0467
www.intrans.iastate.edu
v
TABLE OF CONTENTS
ACKNOWLEDGMENTS ............................................................................................................ vii
INTRODUCTION ...........................................................................................................................1
COST IMPACTS .............................................................................................................................3
Phase One.............................................................................................................................3 Phase Two ............................................................................................................................5 Phase Three ..........................................................................................................................6
IMPACT ON RAIL FREIGHT DEMAND .....................................................................................8
PUBLIC POLICY ISSUES ...........................................................................................................14
REFERENCES ..............................................................................................................................17
vi
LIST OF FIGURES
Figure 1. Estimated costs for Phase One .........................................................................................4 Figure 2. Estimated costs for Phase Two .........................................................................................6
Figure 3. Estimated costs for Phase Three .......................................................................................7 Figure 4. Histogram and density plot of shipment distances by mode ..........................................10 Figure 5. Graphical illustrations of equations ................................................................................11
LIST OF TABLES
Table 1. Summary of effects on truck cost per mile ........................................................................7
Table 2. Percent decrease in rail demand.........................................................................................9 Table 3. Ad hoc method: Percent change in rail demand (Q) ........................................................13
vii
ACKNOWLEDGMENTS
The authors would like to thank the University of Missouri-St. Louis, the Midwest
Transportation Center, and the U.S. Department of Transportation Office of the Assistant
Secretary for Research and Technology for sponsoring this research.
1
INTRODUCTION
The United States has taken great measures to successfully develop its extensive transportation
networks among all modes. There were the railroad subsidies of the late 19th and early 20th
centuries, massive highway projects funded by the Federal government in the mid-1900s, as well
as the airport development grants at about the same time. More recently, there was the
Intermodal Surface Transportation Efficiency Act of 1991, which sought to improve highway
and rail corridors that were deemed critical for facilitating intermodal transportation.
Governments all around the world are looking toward intermodalism to address the problem of
increasing transportation demands per capita, which has been on an upward trend for decades
(Fagnant and Kockelman 2015). Given the fact that an expansion of the existing transportation
infrastructure for any given mode comes at a high cost, intermodal transport has become the
answer for many companies and thus for the government as well.
However, new and emerging technology has the potential to be a disruptive force in these efforts.
For instance, further government assistance in the intermodal transportation industry may prove
to be a frivolous endeavor in the face of driverless vehicles. Driverless vehicles can drastically
reduce the private cost of truck freight shipments by an amount equal to the cost of labor.
Furthermore, if these vehicles are operated in a “platoon,” then there are potential fuel savings
that can further reduce the cost of truck freight. Higher fuel prices lead to higher marginal fuel
cost savings, and with possible total cost savings of over one third, it seems reasonable to
question whether further subsidies towards intermodal transportation is as necessary as before.
Truck platooning involves at least two trucks driving in a perfectly synchronized fashion aided
by computer technology. Platooning requires trucks to drive in tandem such that they maintain a
small distance apart for the air flow around both trucks to work synergistically to reduce drag for
both the end vehicle(s) and the front vehicle; thus, it saves fuel for all those involved. According
to recent studies, platooning can have anywhere from 2% to 15% fuel savings depending on the
gap between trucks, with the fuel savings to be negatively related to the distance gap (Dávila and
Nombela 2011). The basic idea is to have one driver at the head of the platoon or convoy, and
the vehicles that are following closely behind are completely controlled by software that mimics
the lead vehicle to maintain a specific distance from one another in a synchronized fashion.
Whether the following vehicles will house drivers depends entirely on the response of legislative
authorities and technological development.
Emerging automated driving technology can provide real and significant benefits to society,
especially the trucking industry. However, there are many who fear the implications of
widespread use of such technology in the private sector and, in particular, the commercial sector.
These fears are not unreasonable nor without basis. New technology, such as autonomous
driving software, leaves itself open to many vulnerabilities as it tries to mimic the complex
decision-making process that occurs in humans despite having the advantageous 360 degree
sensor range and a seemingly eternal attention span.
Most of these technologies employ programs that are capable of “self-learning,” otherwise
known as machine learning or artificial intelligence. However, as promising as these
2
technologies sound, they are only able to achieve functioning status after a certain amount of
“experience.” These programs require a massive amount of training data from which the program
is supposed to learn and thus appropriately respond to a whole host of potential traffic situations.
Furthermore, this kind of technology often requires continuous global positioning system (GPS)
tracking that is open to software attacks or hacks. If an autonomous vehicle (AV) were involved
in an accident, there would be serious issues regarding accountability. These are the primary
issues that concern lawmakers and provide justification for postulating multiple phases to
characterize the legal progression associated with the employment of such technologies.
Both autonomous vehicle and connected vehicle (CV) technologies have direct benefits to the
motor carrier industry because their implementation would translate into significant levels of cost
reduction. But to evaluate the cost impacts, it is important to understand that these effects will
vary with the technology and how it is permeated and regulated. Therefore, three phases were
posited to characterize a possible progression in how AV will be deployed if CV capabilities are
available throughout all stages.
The focus of these analyses is the potential effects of the technology (at each stage of adoption)
on the cost of truck freight relative to that of rail intermodal services. The first chapter introduces
the three phases as well as their effects on truck costs and breakeven points between trucks and
rail. The next chapter explores the potential effects on the rail freight market. The third and final
chapter lays out some of the study limitations and policy implications.
3
COST IMPACTS
The three phases presented here differ by available AV and CV technology, as reflected by the
legislative climate. AV technological advancements have a standard characterization provided by
the Society of Automotive Engineers (SAE), but CV technology does not. Therefore, it was
necessary to be explicit about how CV technology differs between the posited phases. The only
CV application considered in this report is platooning, which is arguably the most relevant CV
application within the motor carrier industry. Since platooning can have any number of trucks
greater than one, then, for simplicity, only platoons of two trucks were examined.
Phase One
This initial stage is characterized by the existence of regulation that prevents any truck in a
platoon to operate without a human driver, which means that SAE Level 3 is the highest
available level of autonomy. Though it is correct to point out that the highest level of autonomy
that would be on the road also depends on technological advancements, it is useful to view
legislative developments as a function of technological capabilities. Thus, it is sufficient to
consider legislation as the only determinant in the allowable degree of autonomy.
In this phase, AV technology above SAE Level 3 was not yet viable, and CV technology exists
such that the following trucks within a platoon mirrored the leading truck in real time. Here, the
ability for a motor carrier to reduce its labor force existed in so far as there were more employees
than there were trucks. Typically, motor carriers want to utilize their trucks as much as possible
and avoid any idle time to maximize returns on capital investments. Due to the maximum
consecutive work hours imposed by the Federal Motor Carrier Safety Administration (FMCSA),
as well as biological limitations, carriers are forced to have a labor force larger than the stock of
vehicles. Drivers take turns throughout the week, but the same vehicle is utilized.
A platoon scheme can reduce the need for a large labor force since the “drivers” in the following
vehicles do not have to do any driving while the vehicle is in transit. This allows for carriers to
decrease their labor force since one driver could effectively utilize a vehicle for much longer.
However, carriers will still not be able to substitute away from the labor required per unit of
“output,” since there are limits to how long drivers can be “on duty,” which includes time spent
not operating a vehicle. That means that the benefit from logging off-duty hours while in the
following vehicle maxes out at 3 hours due to hours of service (HOS) regulations, thus enforcing
a maximum of 14 consecutive “on-duty” hours.
Furthermore, drivers are generally paid on a per mile basis, which means that if a carrier can
reduce the size of its labor force, it would not matter since it will still drive the same number of
miles (i.e., the man-hours per unit of output will be the same). Therefore, the most important
issue that is addressed in this section is whether companies will still find it profitable to invest in
the technology even if they cannot substitute away from labor.
4
The average cost per mile in the motor carrier industry during 2015 was $1.593, with fuel
making up 25% of that cost and driver wages and benefits accounting for about 39.5% (Torrey
and Murray 2016). Given that fuel is the second highest variable cost that trucking firms face,
platooning technology could still offer substantial savings. However, there is a significant caveat
to this. Since a platoon requires at least two trucks, then any cost savings that would result from
the implementation of a platoon would also require that there be at least two shipments going in
the same direction at the same time. Thus, it is assumed that third party logistics (3PLs) and
freight forwarding companies can match shipment demands so that they can capture the benefits
of platooning technology. Likewise, the technology could also make it possible for trucks to
match with one another in real time in a decentralized fashion, since they often drive alongside
one another. Therefore, it is assumed that platooning technology can be utilized for every long-
haul shipment for simplicity and because the focus of this study was to analyze the potential
effects of platoons.
The first step included graphing the relative cost curves and breakeven point between the two
modes under consideration. Per Boardman et al. (1999), the estimated cost per container mile for
rail was $0.35 per container mile with a loading fee of about $250. The Producer Price Index
(PPI) from Research and Innovative Technology Administration (RITA) was used to put these
cost estimates into 2015 dollars, which gave rise to Figure 1.
Figure 1. Estimated costs for Phase One
The truck and rail curve shows that the 2015 estimated breakeven point was somewhere in the
ballpark of 455 miles. This meant that if there were no other preferences outside of price, then a
consumer would be indifferent between using rail and truck for any shipment with a distance
around the 455 mile mark, and rail would have a price advantage above the 455 mile mark.
The cost per mile and fuel cost per mile for trucks used in these calculations was $1.593 and
$0.403, respectively (Torrey and Murray 2016). The platoon maximum and minimum curves
5
correspond to the cost per mile at the maximum and minimum possible fuel savings that would
result from platooning. As Figure 1 indicates, platooning increases the breakeven distance
between truck and rail by a small amount (somewhere between 11 to 21 miles). This is a result of
the fact that a platoon at this phase has the potential to reduce the average fuel cost per mile from
anywhere between 5 to 12%, which translates to a reduction in the total cost per mile by about
1% to 3%.
Phase Two
A reasonable next step in legislation regarding automated driving technology and platoons would
be the allowance of an effective driver-to-truck ratio of less than one but more than zero. This
phase does not necessarily require higher AV capabilities but rather higher CV capabilities, since
the following vehicles that lack drivers are not acting independently of human control. Thus, this
phase requires CV technology to be advanced enough to not only mirror the lead vehicle but to
perfectly time the mirroring so that making turns and lane changes is safe. Under a two-truck
platoon scheme, this means that the only allowable configurations are two trucks/two drivers or
two trucks/one driver. The following equation details the calculation that describes the labor cost
savings of shipping 𝑛 truck containers from A to B:
𝑆𝐴𝐵=𝐷𝐴𝐵×𝑛𝐴𝐵× [𝐶𝐴𝐵− (1−𝑅) ×𝐶𝐵𝐿] (1)
𝑆𝐴𝐵 = Cost savings of labor reduction in dollars for a shipment between point A to B 𝑛𝐴𝐵 = Number of containers going from A to B with similar time schedule 𝐶𝐴𝐵 = Cost per container going from A to B 𝐶𝐴𝐵𝐿 = Cost per unit of labor going from A to B 𝐷𝐴𝐵 = Distance between AB in miles 𝑅 = Legally allowed driver to truck ratio
For simplicity and without loss of generality, the framework presented here will focus on a two-
truck platoon schematic, i.e., 𝑛𝐴𝐵=2 is assumed.
Further, it is easy to see that this assumption restricts the value for 𝑅 to 0.5. A wage per mile of
$0.499 is added together with the driver benefits per mile of $0.131 to obtain a total labor cost
per mile of $0.63. Using Equation 1 and normalizing 𝐷𝐴𝐵 to equal 1, then the labor cost savings
per mile per container is about $0.32. This translates to a new reduced cost per mile of about
$1.28, and, after including the fuel savings, which ranges from 2.97 to 1.25%, a final estimated
reduced cost per mile is obtained that is between $1.24 and $1.26. Thus, depending on the gap
distance between trucks, a platoon could result in a reduction in the cost per mile between 21%
and 22.3% (see Figure 2).
6
Figure 2. Estimated costs for Phase Two
Phase Three
This phase is characterized by legislation that allows for an autonomous platoon (i.e., 𝑅 = 0 is
allowed). For this legislation to materialize, SAE Level 4 or higher is necessary. The analysis in
this section is like that of Phase Two but is greatly simplified. Using the $1.593 cost per truck
mile figure and subtracting the total labor cost per mile of $0.63 gives $0.963 as the new total
cost per mile. Applying the fuel cost savings gives an estimated cost per mile for trucks ranging
from $0.923 to $0.943.
Figure 3 illustrates the enormous increase in the breakeven point that would result from an
autonomous platoon operation. In all, a full reduction in driver cost and a partial reduction in fuel
cost amounts to about a 40% decrease in the cost per mile.
7
Figure 3. Estimated costs for Phase Three
The first two columns of Table 1 include the estimated new cost per mile if a platoon were
implemented with an SAE Level 4 or higher system.
Table 1. Summary of effects on truck cost per mile
Phases
Maximum Fuel
Savings
Minimum Fuel
Savings
Percent
Decrease Max.
Fuel Savings
Percent
Decrease Min.
Fuel Savings
Phase One $1.55 $1.57 2.7% 1.4%
Phase Two $1.24 $1.26 22.1% 20.9%
Phase Three $0.923 $0.943 42.1% 40.8%
The third and fourth columns of Table 1 include the percent decrease in cost per mile associated
with the values in the first two columns. It is also easy to envision a motor carrier industry that
primarily uses autonomous vehicle technology rather than platooning. Without the added
benefits of platooning, the industry could still enjoy a substantial potential reduction in costs per
mile of about 39.6%.
8
IMPACT ON RAIL FREIGHT DEMAND
AV and CV (platooning) technologies can essentially be viewed as cost-lowering technologies
for long-haul truck freight services. Furthermore, long haul truck freight services are often seen
as a direct substitute to rail freight services, whereas short-haul truck freight can be viewed as a
compliment to rail. Since platooning is only economical over long distances and AV technology
is likely to be first used on Interstates and highways, this research treated the two modes (truck
and rail) as substitutes. As such, the typical way of evaluating the impact of a price drop of one
service on the demand of a separate service requires the use of an estimated cross-price elasticity
of demand.
Abdelwahab (1998) estimated modal elasticities for freight demand using a three-equation
simultaneous discrete choice modelling technique with a disaggregated dataset. The study
estimated the same model separately for different commodity types, and thus each commodity
had different estimated cross-price elasticities. The findings were aggregated in this study to
evaluate the total effect on the rail industry. Data from the 2012 Commodity Flow Survey Public
Use Microdata was used, which contained data on shipment characteristics (U.S. Census Bureau
2015). First, observations of rail-only modes were selected and assigned a value for the estimated
cross-price elasticity, the value of which depended on commodity type. If the specific
commodity type in the dataset could not be categorized as any of those from Abdelwahab (1998),
then it was assigned a value equal to the arithmetic mean of all commodity types. Since the
cross-price elasticity of demand for rail with respect to truck is given by
(2)
where 𝑞𝑖𝑅 = ton-miles for rail and shipment 𝑖, rearranging the terms then gives the estimated
market response for a given shipment by
(3)
This equation is then applied to the dataset using the percent change in truck costs from Table 1
as the percent change in truck prices. After calculating %Δ𝑞𝑖𝑅 for each row in the data, it is then
multiplied by the shipment tabulation weighting factor, which is an estimate of the number of
shipments that are represented by a unique row:
(4)
9
𝑄𝑅=Σ𝑤𝑓𝑖×𝑑𝑖×𝑤𝑔ℎ𝑡𝑖𝑖 𝑤𝑔ℎ𝑡𝑖= shipment weight for shipment 𝑖 𝑤𝑓𝑖= weighting factor for shipment 𝑖 𝑑𝑖= route distance for shipment 𝑖
Table 2 presents the estimated impacts on the total rail freight market based on the results from
Abdelwahab (1998) and Equations (3) and (4).
Table 2. Percent decrease in rail demand
Phases
Percent Decrease in Q
(Max. Fuel Savings)
Percent Decrease in Q
(Min. Fuel Savings)
Phase One 2.91% 1.56%
Phase Two 23.89% 22.54%
Phase Three 45.34% 43.99%
Note: “Q” represents the quantity demanded for rail freight transportation services
A potential limitation of these results is that they are based on data from 1981, which is almost
40 years ago. It is also entirely possible that the underlying structure of the freight transportation
industry has changed sufficiently to hinder the usefulness of these elasticity estimates. Thus, it is
useful to cross-validate these results with a more current data set. However, instead of using the
methods presented in Abdelwahab (1998), an alternative ad hoc method of estimating the rail
market response is presented, which avoids the need of running simultaneous discrete choice
models.
This alternative method of estimating the percent decrease in rail demand relies on the accuracy
of the cost curves and strict distributional assumptions about shippers’ value functions. It is
assumed that the breakeven point given by the intersection of the two curves is correct, or at least
sufficiently close.
Figure 4 clearly shows that there exist rail shipments that lie below (to the left of) the projected
breakeven point (illustrated by the black vertical line) between rail and truck (without platoon)
(as shown in Figure 1).
10
Figure 4. Histogram and density plot of shipment distances by mode
This is easily explained by introducing other factors that affect the decision of modal choice,
such as time (t), consistency (c), flexibility (f), reliability (r), and accessibility (a) of service.
Therefore, if a shipment with a distance value that is to the left of the breakeven point, B, is made
over rail, then the following must be true:
𝑉𝑇𝐴 (𝑐,𝑓,𝑟,𝑎)− 𝐶𝑇− [𝐶𝑡 ×𝑡𝑇] ≤ 𝑉𝑅
𝐴 (𝑐, 𝑓, 𝑟, 𝑎) −𝐶𝑅− [𝐶𝑡×𝑡𝑅] (5)
𝐶𝑅−𝐶𝑇≤ (𝑉𝑅𝐴−𝑉𝑇
𝐴) −𝐶𝑡 (𝑡𝑅−𝑡𝑇) (6)
𝑉𝑇𝐴 (𝑐,𝑓,𝑟,𝑎) = A function that gives the dollar value of all the nonpecuniary attributes for truck
freight services; is a function of consistency (c), flexibility (f), reliability (r), and accessibility of
service (a)
𝑉𝑅𝐴 (𝑐,𝑓,𝑟,𝑎) = A function that gives dollar value of all nonpecuniary attributes for rail freight
𝐶𝑡 = Dollar cost of a unit of travel time
𝑡𝑇 = Time required for truck to complete service
𝑡𝑅 = Time required for rail to complete service
𝐶𝑇 = Cost of truck service, calculated as the cost per mile multiplied by distance
𝐶𝑅 = Cost of rail service, calculated as the cost per mile multiplied by distance plus the loading
fees
Thus, the difference in values of nonpecuniary service attributes minus the cost of additional
travel time has a lower bound given by the vertical distance between the curves Truck_0 and Rail
in Figure 5 (e.g., the value labelled as pd_2).
11
Figure 5. Graphical illustrations of equations
Intuitively, what this says is that if the estimated cost curves for truck and rail are accurately
represented, then the only way that a consumer would choose rail over truck for a shipment
distance that lies below B is if the value of the nonpecuniary service attributes of rail is
sufficiently greater than that of truck, such that it more than compensates for the dollar difference
in cost.
To get an upper bound for the implicit value difference, however, it is assumed that if the
pecuniary cost of truck were zero everywhere, then consumers would choose truck over rail
regardless of the implicit value difference of attributes. Thus, the final inequality is given:
𝐶𝑅−𝐶𝑇≤ (𝑉𝑅𝐴−𝑉𝑇
𝐴) −𝐶𝑡 (𝑡𝑅−𝑡𝑇) ≤𝐶𝑅 (7)
Then, by comparing the net change in price differences for each unique distance that would
result from platooning, it is possible to see whether the cost savings are enough to compensate
for the implicit value difference. Applying this to each unique distance value, it is possible to
estimate the proportion of demand that lies below B that would switch to truck.
Let 𝑓(𝑑𝑖) be the cost curve for truck services, 𝑔(𝑑𝑖) be the cost curve for platoon services, and 𝑑𝑖 be the unique distance value for shipments that is observed in the CFS data with 𝑑𝑖∈(0,𝐵). Only
the data for rail shipments are used here, and the following is calculated:
(𝑑𝑖)−𝑔(𝑑𝑖) (8)
The observations (with observations each being 𝑑𝑖), which satisfies the inequality, are kept as
12
𝐶𝑅−𝐶𝑇≤ (𝑑𝑖)−𝑔(𝑑𝑖) (9)
For each 𝑑𝑖, there are 𝑤𝑓𝑖 numbers of shipments, and we assume that these shipments can have
varying levels of implicit value differences of service attributes that are uniformly distributed
throughout the range defined by Equation (7). This then allows for the estimation of the
proportion of shipments, for a given 𝑑𝑖, that would switch from rail to truck, and is given by
(10)
Adding these all up gives the total demand measured as ton-miles, which lies below B, and will
switch over to truck after the price change.
(11)
where 𝜃𝑖 = 𝑑𝑖∗𝑤𝑓𝑖∗𝑤𝑔ℎ𝑡𝑖 and 𝑤𝑔ℎ𝑡𝑖 denotes the weight in pounds of a shipment. For those rail
shipments that lie to the right of B, Equation (5) must also hold. However, being to the right of B
means that 𝐶𝑇 > 𝐶𝑅, which gives
(12)
Here, it is assumed that 𝑉𝑇𝐴 (𝑐, 𝑓, 𝑟, 𝑎) > 𝑉𝑅
𝐴 (𝑐, 𝑓, 𝑟, 𝑎), because if the opposite were analyzed, then
it would be uninformative since the cost of rail is also lower than that of truck for distances beyond
B. The fact that consumers choose rail would tell us nothing about how the implicit value differences
are bounded.
(13)
Similarly, the following gives the number of ton-miles to the right of B that will switch over to
truck after the price change, where ℎ(𝑑𝑖) is the cost curve for rail and D denotes the highest
distance value observed in the data.
(14)
D is the distance at which ℎ(𝐷)=𝑔(𝐷).
13
Adding the two summation equations together gives the total demand that will switch to trucks
after a price change. Then, dividing the total demand that will switch by the total number of ton-
miles on both sides of B gives the estimated percent change total rail demand. These estimates
are presented in Table 3 for all three phases.
Table 3. Ad hoc method: Percent change in rail demand (Q)
Phases
Percent Decrease in Q Max.
Fuel Savings
Percent Decrease in Q Min.
Fuel Savings
Phase One 0.51% 0.2%
Phase Two 5.93% 5.44%
Phase Three 20.4% 18.59%
14
PUBLIC POLICY ISSUES
The cross-price elasticity estimates that result from the ad hoc approach is much less than unity,
and those that are used by Abdelwahab (1998) are just a bit above unity. Abstracting from
technical considerations for a moment, it is possible to conceive a story that can explain why rail
freight demand is either responsive or unresponsive to truck prices. On the one hand, it is
generally believed that the attributes associated with truck freight services are valued more than
those of rail on almost every dimension other than accessibility, which probably depends on the
nearby infrastructure and is idiosyncratic to any given shipping scenario. So, if shippers are
choosing rail over truck even though truck costs are lower and service attributes are better than
rail, then it must be due to accessibility reasons. Thus, it can be argued that rail freight demand is
highly responsive to truck freight prices because, up to a certain distance, truck outcompetes rail
so much so that consumers of rail services are only marginally attached.
The flipside of this is that, for similar reasons, it can also be argued that the value associated with
rail service accessibility must be extremely high to compensate for the many disadvantages of
rail and results in a very unresponsive rail demand with respect to truck prices. The stark
differences between the estimates from the two methods highlight the need for a more up-to-date
estimation of the cross-price elasticity of rail freight demand relative to truck prices.
One of the downsides to the ad hoc method that was used in the previous section is that it relies
on several convenient and strict assumptions about preferences. This limits the usefulness for
policy decisions. However, the results from Abdelwahab (1998) are also limited in its usefulness
considering that the cost structures of the two modes have changed significantly. Over the last 20
years or so, the producer price index for line-haul rail operations increased much faster than that
of long-haul trucking, which suggests that the elasticities from Abdelwahab (1998) may not
necessarily hold (Bureau of Transportation Statistics 2018). Therefore, there are no obvious ways
of determining which of the two methods dominate.
With the rapid pace of AV/CV technological development, it is likely that fully automated trucks
will be in use in half this time. Therefore, a portion of the rail intermodal funds would be vastly
underutilized, as well as the associated costs from the decision-making process used to allocate
these funds.
In many cases, the process of applying for and receiving federal funds for local transportation
infrastructure projects are costly, in both time and money, for all those involved. This type of
undertaking has varying rates of success as well, because a project requires unanimous
agreement from all stakeholders that would be affected by the completion of the proposed
project. In contrast, Interstate projects are controlled by the federal government, and any decision
to improve Interstate infrastructure to accommodate new motor carrier technologies would not
have as high of an administrative cost.
In evaluating whether any given infrastructure project is worth funding, it is useful to assess the
payback period of the project in relation to the development and adoption stages of platooning
and driverless truck technologies. It is also important to understand the implications that these
15
stages would have on the overall freight network. The previous sections have shown that the
implications of Phase One on the rail freight demand are small but grow larger as the phase
progresses. In response to this, the rail industry could conceivably bring down the fixed loading
costs (intercept of the rail cost curve) of rail and make the cost of rail at least as competitive as
that of truck after platoon adoption.
Moreover, if the National Freight Strategic Plan sets the goal of having more long-distance
freight be moved by rail, using primarily intermodal services, then the only way that the
researchers can suggest this be done is for the federal government to purchase and maintain the
rail infrastructure to help lower the fixed cost of rail operations. Railroads would still pay a
variable fee per ton-mile of track utilized, similar to the fuel fees paid by the trucking industry
for the use of public highways. However, their operating costs could then compete with those of
the motor carrier industry.
Despite this, there are still a multitude of factors that work against the rail industry and
subsequently the demand for intermodal freight service. The proliferation of just-in-time delivery
as an inventory strategy creates a market that favors a delivery service with high scheduling
flexibility and time consistency. Both of these are advantages that trucks have over rail, and this
discrepancy will only increase with the advent of the new trucking technologies considered in
this report.
It is ironic that our nation’s first railroads were run on public right-of-ways. Anyone could run on
these primitive wooden rails, staying out of the mud, for a fee. Perhaps this may be the only way
for public policy to lessen congestion on our highways of the future.
17
REFERENCES
Abdelwahab, W. M. 1998. Elasticities of mode choice probabilities and market elasticities of
demand: evidence from a simultaneous mode choice/shipment-size freight transport
model. Transportation Research Part E: Logistics and Transportation Review, Vol. 34,
No. 4, pp. 257–266.
Boardman, B. S, E. M. Malstrom, and K. Trusty. 1999. Intermodal Transportation Cost Analysis
Tables. Mack-Blackwell National Rural Transportation Study Center, University of
Arkansas, Fayetteville, AR.
Bureau of Transportation Statistics. 2018. Table 3-13: Producer Price Indices for Selected
Transportation and Warehousing Services.
https://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportatio
n_statistics/html/table_03_13.html.
Dávila, A. and M. Nombela. 2011. Sartre-safe Road Trains for the Environment Reducing Fuel
Consumption through Lower Aerodynamic Drag Coefficient. SAE International,
Warrendale, PA.
Fagnant, D. J. and K. Kockelman. 2015. Preparing a nation for autonomous vehicles:
opportunities, barriers and policy recommendations. Transportation Research Part A:
Policy and Practice, Vol. 77, pp. 167–181.
Torrey, W. F. and D. Murray. 2016. An Analysis of the Operational Costs of Trucking: 2016
Update. American Transportation Research Institute, Arlington, VA. http://atri-
online.org/wp-content/uploads/2016/10/ATRI-Operational-Costs-of-Trucking-2016-09-
2016.pdf.
U.S. Census Bureau. 2015. 2012 Commodity Flow Survey Public Use Microdata.
https://www.census.gov/econ/cfs/pums.html.
Visit www.InTrans.iastate.edu for color pdfs of this and other research reports.
THE INSTITUTE FOR TRANSPORTATION IS THE FOCAL POINT FOR TRANSPORTATION AT IOWA STATE UNIVERSITY.
InTrans centers and programs perform transportation research and provide technology transfer services for government agencies and private companies;
InTrans manages its own education program for transportation students and provides K-12 resources; and
InTrans conducts local, regional, and national transportation services and continuing education programs.